Clinical Research Collaboration for Stroke in Korea Imaging Repository:A Prospective Multicenter Neuroimaging Repository

Technical Deep Dive: Clinical Research Collaboration for Stroke in Korea Imaging Repository

Overview

This study describes the development and initial findings from the Clinical Research Collaboration for Stroke in Korea (CRCS-K) Imaging Repository. This is a prospective multicenter platform that systematically collects and analyzes comprehensive stroke neuroimaging data, integrating it with clinical and outcome data through an AI-powered research platform called AISCAN.

Problem & Context

Stroke is a major cause of disability and mortality worldwide. While prospective stroke registries have advanced our understanding of cerebrovascular disease, most reduce neuroimaging data to categorical variables, missing the nuanced, multidimensional information inherent in clinical imaging. The CRCS-K Imaging Repository aims to address this gap by systematically collecting all stroke neuroimaging (CT, MRI, angiography) and applying AI-based automated quantification to extract standardized numeric features.

Methodology

• The CRCS-K Imaging Repository was built upon the existing CRCS-K, a nationwide prospective stroke registry in Korea.
• All neuroimaging (CT, MRI, angiography) performed during the index hospitalization of consecutive acute ischemic stroke patients was collected from 18 comprehensive stroke centers.
• The imaging underwent centralized quality verification, sequence classification, and AI-based quantification to extract standardized numeric features such as ischemic lesion volumes, perfusion parameters, white matter hyperintensity burden, and cerebral microbleed counts.

Data & Experimental Setup

• From June 2022 through May 2025, the repository collected 225,159 imaging sequences from 20,792 patients.
• As a proof-of-concept analysis, the researchers examined the association between pre-treatment imaging modality (CT vs. MRI), treatment workflow efficiency, and functional outcomes in patients receiving intravenous thrombolysis (IVT) or endovascular treatment (EVT).

Results

• Substantial inter-hospital variation was observed in imaging modality selection, with MRI-first workflows ranging from 1.0% to 56.7% across centers.
• Each additional imaging sequence was associated with prolonged door-to-treatment times for both IVT and EVT.
• Propensity score analyses suggested numerically more favorable functional outcomes with CT-based imaging among EVT-treated patients, whereas differences among IVT-treated patients were smaller and less consistent.

Interpretation

The CRCS-K Imaging Repository demonstrates the feasibility of large-scale, prospective neuroimaging collection integrated with AI-based quantification and clinical data. This infrastructure enables clinically consequential questions that conventional registries cannot address, such as the impact of imaging modality selection on treatment workflow and patient outcomes.

Limitations & Uncertainties

• The study relied on observational data, so causality cannot be firmly established.
• The proof-of-concept analysis was limited to a specific treatment context (IVT and EVT) and may not generalize to other stroke management approaches.
• The AI-based quantification, while standardizing the imaging data, may still be subject to inherent biases and uncertainties.

What Comes Next

The CRCS-K Imaging Repository provides a powerful platform for addressing a wide range of clinically relevant questions in stroke care, from optimizing imaging-based patient selection to evaluating novel imaging biomarkers for prognosis and treatment guidance. Future research will leverage this infrastructure to explore these and other important topics in stroke neuroscience.